Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
NBER WORKING PAPER SERIES
SAFETY-NET BENEFITS CONFERRED ON DIFFICULT-TO-FAIL-AND-UNWINDBANKS IN THE US AND EU BEFORE AND DURING THE GREAT RECESSION
Santiago Carbo-ValverdeEdward J. Kane
Francisco Rodriguez-Fernandez
Working Paper 16787http://www.nber.org/papers/w16787
NATIONAL BUREAU OF ECONOMIC RESEARCH1050 Massachusetts Avenue
Cambridge, MA 02138February 2011
The authors thank the Fundación de las Cajas de Ahorros (Funcas) for supporting this research. Wealso thank James Wilcox, Robert Eisenbeis and other participants in the 2011 ASSA meetings andfrom Marianne Verdier and other participants in the seminar held at the Université Paris Ouest-Nanterrela Défense in January 2011. Comments from Ethan Cohen-Cole, Robert Dickler, Stephen Kane, andJames Thomson are also acknowledged and appreciated. Santiago Carbó and Francisco Rodriguezacknowledge financial support from the Spanish Ministry of Science and Innovation and FEDER (ECO2008-05243/ECON) and from the Consejería de Innovación, Ciencia y Empresa-Junta de Andalucía (P08-SEJ-03781). Kane is grateful for support received on a related project from the Institute for New EconomicThinking. The views expressed herein are those of the authors and do not necessarily reflect the viewsof the National Bureau of Economic Research.
NBER working papers are circulated for discussion and comment purposes. They have not been peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies officialNBER publications.
© 2011 by Santiago Carbo-Valverde, Edward J. Kane, and Francisco Rodriguez-Fernandez. All rightsreserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permissionprovided that full credit, including © notice, is given to the source.
Safety-Net Benefits Conferred on Difficult-to-Fail-and-Unwind Banks in the US and EU Beforeand During the Great RecessionSantiago Carbo-Valverde, Edward J. Kane, and Francisco Rodriguez-FernandezNBER Working Paper No. 16787February 2011JEL No. G01,G2,G21,G28,G38,K2
ABSTRACT
This paper models and estimates ex ante safety-net benefits at a sample of large banks in US and Europeduring 2003-2008. Our results suggest that difficult-to-fail and unwind (DFU) banks enjoyed substantiallyhigher ex ante benefits than other institutions. Safety-net benefits prove significantly larger for DFUfirms in Europe and bailout decisions less driven by asset size than in the US. We also find that a proxyfor regulatory capture helps to explain bailout decisions in Europe. A policy implication of our findingsis that authorities could better contain safety-net benefits if they refocused their information systemson measuring volatility as well as capital.
Santiago Carbo-ValverdeDepartamento de Teoría e Historia EconómicaFacultad de CCEE y EmpresarialesUniversidad de Granadas/n E-18071, Granada, [email protected]
Edward J. KaneDepartment of FinanceBoston CollegeChestnut Hill, MA 02467and [email protected]
Francisco Rodriguez-FernandezDepartamento de Teoría e Historia EconómicaFacultad de CCEE y EmpresarialesUniversidad de Granadas/n E-18071, Granada, [email protected]
February 1, 2011
2
SAFETY‐NET BENEFITS CONFERRED ON DIFFICULT‐TO‐FAIL‐AND‐UNWIND BANKS IN THE US AND EU BEFORE AND DURING THE GREAT RECESSION
Santiago Carbo‐Valverde Edward J. Kane
Francisco Rodriguez‐Fernandez
A nation’s financial safety net is a set of programs aimed at protecting bank depositors and
keeping systemically important markets and institutions from breaking down in difficult circumstances.
Although different instruments and functions are often located in different agencies, the organization
can be seen as a government‐owned holding company. Considered as a consolidated enterprise, this
collection of agencies has a balance sheet, an income statement, and a governance network across
which stakeholders and managers interact. Its governance procedures are complicated by differences in
the capacities of different stakeholders to understand and promote their interests and these differences
vary widely across countries.
The current financial crisis has hit some countries much harder than others. This paper seeks to
benchmark differences in how well, both before and during the current crisis, safety‐net managers in
the US and 14 European countries managed the tradeoff in their systems of institutional support
between the interests of bankers and taxpayers.
The principal goal of safety‐net management is to monitor, contain, and finance systemic risk.
Systemic risk combines two kinds of risk‐taking: calculated risk‐taking by protected institutions and
partially countervailing risk‐management programs operated by safety‐net managers. Ideally, safety‐net
managers safeguard taxpayers interests by making institutions operate more safely than stockholders
might prefer.
Definitions of systemic risk used by the Basel Committee and other policymakers focus on a
perceived potential for substantial spillovers of institutional defaults across important firms in the
financial sector and from this sector to employment and asset values in the real economy. But this
February 1, 2011
3
perspective conceives of safety‐net managers as managing a negative externality rather than completing
markets by tolerating and redistributing downside risk to taxpayers. Because research indicates that
stock markets do price downside risk at protected institutions, taxpayer exposure to this risk is not
strictly an externality. The conjectural contingent obligations that government safety nets impose on
taxpayers guarantee the performance of particular contracts in adverse circumstances. In effect, they
complete what would otherwise be a system of incomplete markets for the liabilities of protected firms
(cf., Kane, 1980). Since costs to taxpayers of underwriting tail risks at these institutions are not expected
to be recovered in full by means of ex post assessments, national safety nets are programs of
redistributive fiscal policy that encourage excessive risk taking by systemically important firms.
One lesson of the current crisis is that the aggregate costs and benefits of producing
government guarantees are not much different from their private costs and benefits. Despite the global
extent of the current crisis, actual spillovers of defaults have been minimal. Firms that seemed
politically or administratively difficult to fail and unwind (DFU firms) have been characterized as
systematically important and kept running by supporting their access to public and private credit
without resolving their underlying shortage of capital (i.e. their economic insolvency). In effect,
authorities exercised a loss‐shifting “taxpayer put” that converted most of the losses incurred by
insolvent DFU firms into government debt (Kane, 1986; Eberlein and Madan, 2010).
The capitalized value of the safety‐net subsidies that DFU firms capture from taxpayers
represents a cogent way to measure what authorities ought to mean by “systemic risk.” This definition
of systemic risk implies that, even in good times, ex ante safety‐net subsidies exist for DFU firms. Of
course, ex post subsidies become more plainly visible as prominent firms’ booked and unbooked losses
approach a breaking point. Ironically, in economic downturns, this increased transparency can fuel
popular unrest that reduces the flow of ex ante subsidies to banks that do not receive explicit open‐bank
assistance. On average, but not at the margin, we observe such an effect across our sample and
February 1, 2011
4
subsamples. These distribution effects suggest that taxpayer and competitor interests would be better
served‐‐ in good times and in bad‐‐by surfacing estimates of the buildup of safety‐net subsidies and
recording these estimates on the income statements and balance sheets of governments and DFU firms
alike.
Both in Europe and the US, safety‐net managers seek to contain risk‐taking by restricting the
activities of protected institutions and by prescribing risk‐based capital requirements and insurance
premia. The empirical part of this paper uses the Bankscope database and contingent‐claims models of
safety‐net benefits to estimate and compare the value of leverage ratios and ex ante safety‐net benefits
at firms thought or revealed to be DFU in the US and Europe during 2003‐2008. We find that during
both 2003‐2006 and 2007‐2008 DFU banks in both venues enjoyed substantially higher ex ante benefits
than other institutions in the sample. Safety‐net benefits were significantly larger for DFU firms in
Europe, but bailout decisions appear less driven by asset size and more by regulatory capture than in the
United States.
1. Modeling the Determinants of Systemic Risk
Considerable disagreement exists about the best way to measure systemic risk, but everyone
agrees that it arises as a mixture of leverage and the volatility of financial‐institution returns. This paper
employs a two‐equation model developed by Duan, Moreau, and Sealey (DMS, 1992). In this model,
decisions about volatility constrain leverage decisions and, abetted by leverage, generate safety‐net
benefits. The model incorporates the pioneering perspective of Merton (1977, 1978). Adding ideas
from Ronn and Verma (1986) and Hovakimian and Kane (2000), two other studies [Carbo, Kane, and
Rodriguez (2008, 2010)] use this model to undertake cross‐country comparisons of regulatory and
merger policies.
February 1, 2011
5
The DMS model recognizes that market and regulatory discipline force a financial firm to carry
an equity position that outsiders regard as large enough to support the risks it takes. However, creditors
are assumed to regard the conjectural value of the off‐balance‐sheet capital that government
guarantees supply through the taxpayer put as a substitute for on‐balance‐sheet capital supplied by the
firm’s shareholders.
For individual banks, the DMS model consists of a leverage equation and an equation tracking
safety‐net benefits. This model linearizes and slightly expands Merton’s model of deposit insurance
(1977, 1978). Merton portrays safety‐net access as an option that allows bank owners to put the bank
to safety‐net managers in exchange for the face value of the bank’s debt. Ronn and Verma (1986)
suggest scaling down this option’s exercise right to allow for examination lags and political pressures
that make it hard for authorities to enforce their takeover rights.
Firms engage in risk‐shifting whenever they expose creditors, derivatives counterparties or
guarantors to loss without compensating them adequately. The DMS model measures safety‐net
benefits by a variable designated as IPP. IPP is the so‐called "fair insurance premium percentage” per
dollar, per Euro, or per pound of debt that would let taxpayers break even in each period. The model
makes IPP an increasing function of a bank’s asset risk (σV) and leverage. Leverage is measured as the
ratio of the par value of an institution’s debt (B) to the estimated market value of its assets (V).
Duan, Moreau, and Sealey (1992) stress that market and regulatory disciplines prevent value‐
maximizing B/V (leverage) from being chosen independently of σV (volatility). To contain risk‐shifting at
all, counterparties and regulators must require B/V to fall when and as σV increases. The model is only a
quasi‐reduced form because it treats a potentially exogenous variable σV as an exogenous regressor.
This produces a recursive model for bank B/V and IPP:
B/V = α0 + α1σV + ε1 . (1)
IPP = β0 + β1σV + ε2 . (2)
February 1, 2011
6
Equation (1) expresses the idea that regulators and creditors constrain bank risk‐taking to a
mutually acceptable set of leverage and volatility pairs. If safety‐net managers had no incentive conflicts
and could observe σv and control B/V perfectly, they would set B/V so that IPP equaled the value of the
sum of explicit and implicit premiums they could impose on the bank. Taking total derivatives, the slope
coefficients in (1) and (2) may be interpreted as follows:
, (3)
β1 = . (4)
The partial derivatives that appear in equation (4) are positive. They describe the
incremental value that bank stockholders could extract from the safety net if bankers were free to make
unconstrained adjustments in volatility and leverage, respectively. To prevent a corner solution, either
or both of two conditions must be met. Safety‐net officials and private counterparties must monitor
and constrain bank risk taking at the margin or managers must believe that unbridled pursuit of safety‐
net subsidies would work against their career interests. Equations (3) and (4) express the effects of
“outside discipline.” At DFU banks during the years we examine, managerial restraint or “inside
discipline” seems to have been sorely lacking.
Given the external discipline a bank faces, the sign of β1 in equation (2) indicates whether, in
a country’s particular contracting environment and economic circumstances, increases in asset volatility
can increase the value of the implicit and explicit access to safety‐net support that is imbedded in the
bank’s stock price. To neutralize risk‐shifting incentives at the margin, disciplinary penalties that induce
a decline in B/V must be large enough to offset fully whatever increase in IPP might otherwise be
generated by choosing a higher σV. In firms for which the total derivative β1 is positive, marginal risk‐
shifting incentives are not completely neutralized by inside and outside discipline.
For market and regulatory pressure to discipline and potentially to neutralize incremental
risk‐shifting incentives, two conditions must be met:
February 1, 2011
7
Bank capital increases with volatility: α1 < 0,
Guarantee values do not rise with volatility: β1 ≤ 0.
None of the three variables featured in the DMS model is directly observable. However,
Marcus and Shaked (1984) show how to use option‐based models of deposit insurance to track these
variables synthetically. The first step in the Marcus‐Shaked procedure is to obtain tracking values for V
and σV by numerical methods. These values are then used to estimate IPP as the value of a put option
on bank assets (the “default put”). As explained more fully in Hovakimian and Kane (2000), a key step is
to use Ito’s lemma to transform σV into σE, the instantaneous standard deviation of equity returns.
2. Preliminary Look at Mean Sample Experience
Table 1 lists the number of observations in our sample by country. Over a third of the
observations come from the US and Germany and roughly 80 percent come from the last six countries
listed in the table. Table 2 lists the sources from which we obtain the data we analyze. It also introduces
and defines some control and shift variables (such as DFU status) that we incorporate into our
regression experiments. DFU status is proxied ex ante by a size criterion (DFUxa) and ex post by the
receipt of open‐bank assistance (DFUxp).
Table 3 describes the mean behavior of leverage, volatility, and the fair insurance premium
percentage for different groupings of banks. Throughout the paper, regression inputs are calculated in
two different ways: by the Ronn and Verma (RV) procedure and by a maximum‐likelihood (ML) method
developed by Duan (1994). Table 3 also records the results of tests for differences in the mean values
found between US and European banks and between DFU and other banks in various regions. Mean
differences are significant at conventional levels in every instance.
Mean safety‐net benefits range between 10 and 22 basis points. Mean leverage proves
uniformly higher under the ML procedure, while volatility and IPP are often lower. Using either
procedure, both kinds of DFU banks show higher safety‐net benefits than other banks in both regions
February 1, 2011
8
and time frames. In most cases, DFU institutions show more leverage, too. Both before and during the
crisis, DFU banks in Europe show more leverage and safety‐net benefits than DFU banks in the US and
DFUxp banks extracted more benefits than DFUxa firms. During the crisis, DFU banks in Europe and the
US decreased volatility, reduced their leverage and did suffer procyclical cuts in the mean size of ex ante
safety net benefits.
3. Regression Analysis
Difference‐on‐difference regression experiments expand equations (1) and (2) to include three
control variables and three parameter‐shift indicators for DFU banks1. The log of asset size is introduced
as a hard‐to‐interpret proxy that aggregates the influence of political clout, complexity, and public
awareness separately from measures of DFU status per se. Transparency International’s Corruption
Perception Index (10‐CPI) is used to represent cross‐country differences in a government’s susceptibility
to regulatory capture. We include the so‐called “fear index” (VIX) as a way to distinguish the impacts of
marketwide and idiosyncratic volatility.
Pooling precrisis and crisis years, Table 4 applies this model separately to panels of US and
European banks and bank holding companies. The signs of all coefficients and the rough magnitude of p‐
values are similar in all parallel runs.
Given the large size of these samples and the near‐zero value of focal coefficients, the Lindley
Paradox suggests that we employ a more rigorous standard for statistical significance than the
conventional 5 percent. Our discussions benchmark significance at 2 percent, but the reader is free to
adopt a lighter or tougher standard.
The shift variable in the size effect for DFU banks is never significant and is dropped from
subsequent runs. Except for VIX and the corruption index (which proves significant only in Europe where
there is cross‐section as well as time‐series variation), most differences between US and European
1 See Han and Phillips (2011) for a comprehensive discussion of fixed‐effects panel estimation in difference‐on‐difference regression.
February 1, 2011
9
leverage equations meet the significance standard of 2 percent. The effects of asset size on safety‐net
benefits (i.e., on IPP) are similar across countries, but at the margin DFUxa banks in the US extract
slightly more benefits than their European counterparts.
Table 5A shows the effect of employing the ex post definition of the DFU shift variable. In this
experiment, DFUxp banks are banks that received explicit State aid during the crisis. Although R‐squared
remains much the same, this definition renders differences between Europe and the US in coefficients
for idiosyncratic volatility, asset size, corruption, and the intensified role of volatility for DFU banks
sharper and uniformly more significant. In particular, even though DFUxp banks in the US find
themselves penalized more heavily for increased volatility in the leverage equation than DFUxp banks in
the EU, they manage to extract incremental benefits from the safety net more successfully (0.035 in the
US vs. 0.029 in Europe according to the ML procedure). Additionally, the proxy for regulatory capture
(10‐CPI) is significant only for the EU sample.
Table 5B re‐runs the Table 5A regression experiment using Heckman’s (1976, 1978) procedure
for endogenizing the ex post selection process for providing capital support to DFU banks. This
procedure adds a third equation to our model. This selection equation is linked to the other equations
by a variable that Heckman calls Lambda (also known as the Mills Odds Ratio for selection) which is
calculated from the selection model. This linking variable is then added to the list of the potential
determinants of leverage and safety‐net benefits in expanded versions of equations (1) and (2).
Although the value and significance of coefficients in the B/V and IPP models are not much
different from those in Table 5A, differences in the probit selection models for receiving State aid are
markedly different. In Europe, asset size has no significant effect. Idiosyncratic volatility and the
corruption index dominate government bailout decisions in Europe. In particular, the impact of σV on
the probability of receiving State aid in Europe is 0.815 according to the ML model while the impact of
the regulatory capture proxy (CPI‐10) is 0.916. In the US, idiosyncratic volatility is more or less equally
February 1, 2011
10
important, but size has a large effect. The coefficient for size at US banks estimated with either the RV
and ML procedure is roughly 1.50. Limited to time‐series variation, the corruption index shows no
predictive power in the US.
Tables 6A and 6B run the abridged model of Table 4 separately for pre‐crisis and crisis years: i.e.,
for 2003‐2006 and 2007‐2008. The most interesting differences are those in which the subperiod
coefficients both lie substantially above or below those found in the pooled equation. Such a finding
establishes a prima facie case against pooling data across separate regimes. This phenomenon occurs for
the incremental effects on IPP of the DFUxa shift variable in both regions (+), for corruption (+) in
Europe, and for size (‐) and volatility (+) in the US. In particular, taking the ML model as a reference, the
coefficient of the critical DFU shift variable in the IPP equation changes only slightly (from 0.026 to
0.029) in the EU sample from the pre‐crisis to the crisis period. But the coefficient for the US sample
shows increased marginal subsidization, jumping from 0.034 to 0.044. As for the index of corruption
perceptions, the coefficient in the IPP equation almost doubles (from 0.007 to 0.012), while the index
continues to be insignificant in the US sample. The coefficient of σV in the IPP equation increases for the
EU sample from 0.007 in precrisis years to 0.011 in the crisis period while the coefficient for the US
sample increases hardly at all, from 0.013 to 0.015.
Table 7 reports the significance of differences between coefficients in precrisis and crisis years
for US and European banks separately. In Europe, crisis years show an intensification in incremental
subsidization for a few variables and equations: for idiosyncratic volatility on safety‐net benefits; for the
DFU shift variable on leverage under the ML procedure and on IPP using the RV approach; and for
corruption in the leverage equation and in the ML model for IPP. In the US, the incremental effects of
asset size and the DFU shift variable intensify for both variables under both procedures.
Tables 8A and 8B re‐run the experiments of Tables 6A and 6B using the Heckman procedure and
the ex post DFU indicator. For both the precrisis and crisis eras, the signs of all coefficients for the
February 1, 2011
11
European and US samples remain the same. However, the magnitude of individual coefficients is often
reduced. For the precrisis samples, coefficient differences between US and Europe for the corruption
and VIX index are seldom significant, but the greater role for market volatility in explaining US bank
leverage in crisis years continues to be significant.
As in Table 5, the importance of the Heckman experiments lies in creating the opportunity to
examine the selection equations. Asset size (and to a lesser extent, idiosyncratic volatility) is a more
important determinant of bailout assistance in the US than in Europe, while a European country ‘s
corruption index strongly influences its bailout decisions. The inference is that banks may not be too big
to fail in Europe, but they might be too politically connected.
For European and US sample banks, Table 9 shows that the leverage and IPP equations
underwent many statistically significant changes between precrisis and crisis periods. Economically, the
effects on IPP generation are the most interesting. In Europe, the shift in the volatility slope for DFU
banks explicitly receiving State aid increased by roughly 50 percent under both procedures. In the US,
this coefficient also increased, but the effect is smaller and significant only under the RV procedure.
4. Special Cases of Portugal, Ireland, Italy, and Spain
In some European countries affected most severely by the crisis, doubt has arisen about the
government’s ability to resolve the losses experienced by its largest banks. Greece (for which we lack
data), Ireland, Portugal, Italy, and Spain have all seen substantial increases in the credit premium paid
on their sovereign debt. Tables 10A and 10C apply the expanded DMS model to the high‐premium
countries for which we have data.
Although idiosyncratic volatility is always significant in these four countries, market volatility is
not. Time‐series variation in the index of perceived corruption almost always impacts leverage, IPP, and
selection significantly. However, the economic significance of the proxy for susceptibility to regulatory
capture (10‐CPI) is higher in Ireland (0.021 in the ML version of the IPP equation) than in Portugal
February 1, 2011
12
(0.011), Spain (0.008) or Italy (0.006). Size impacts selection except in Portugal. Idiosyncratic volatility
increases safety‐net benefits more in Portugal (0.010) and Ireland (0.018) than in Spain (0.008) and Italy
(0.006).
Table 10C shows that almost all coefficient differences are significant across country pairs.
Ignoring coefficient differences and discarding the market‐volatility term, Table 11 tests for differences
that apply in precrisis and crisis periods when the DMS model fitted to the DFU banks that were bailed
out in these four countries is compared with the Table 6A and 6B models estimated across the full
sample of European banks. The most striking differences between these two panels and periods is the
much greater importance found during the crisis years for asset size and the proxy for susceptibility to
regulatory capture.
5. Lessons and Policy Implications
Three important lessons emerge from our work. The first concerns authorities’ convenient claim
that crisis pressures could not be foreseen. Despite being limited to annual data for key variables,
changes in volatility and leverage consistently help to predict changes in the flow of safety‐net benefits
across different models, regions, and time periods. The second lesson is that the mean flow of ex ante
benefits declined in the face of the increased public accountability generated by the transparency of ex
post bailout expense. Finally, the cross‐country proxy for susceptibility to regulatory capture (the index
of perceived corruption) helps to explain safety‐net benefits and bailout decisions in Europe.
One policy implication of these findings is that authorities could do a better job of controlling
safety‐net benefits if they expanded their information systems so that they could track IPP in a
transparent manner. As intricate as it may seem, the stochastic and econometric plumbing underlying
our equities‐based estimates of volatility and safety‐net benefits is still at an early stage of evolution.
Complementary estimates can be engineered using richer stochastic processes and datasets based on
February 1, 2011
13
the prices of debt and derivative instruments. We encourage others to do this and are confident that
they will.
One way to improve information flow would be to require that bank managers report data on
earnings and net worth more frequently, with civil penalties for fraud and negligent misrepresentation.
Data on market capitalization are available in real time, as are data on stock‐market returns. If the
values of on‐balance‐sheet and off‐balance‐sheet positions were reported weekly or monthly to national
authorities, rolling regression models could be used to estimate changes in the flow of safety‐net
benefits in ways that would allow regulators to observe, manage and report taxpayers’ stake in the
safety net in a more timely manner.
February 1, 2011
14
REFERENCES
Carbo, Santiago, Edward Kane, and Francisco Rodriguez. (2008) “Evidence of Differences in the Effectiveness of Safety‐Net Management in European Union Countries,” Journal of Financial Services Research, 34, 151‐76.
_________________ (2010). "Regulatory Arbitrage in Cross‐Border Mergers Within the EU," Journal of
Money Credit and Banking (forthcoming). Duan, Jin‐Chuan. (1994). “Maximum Likelihood Estimation Using Price Data of the Derivative Contract,”
Mathematical Finance, 4, 155‐67. Duan, J‐C, Arthur F. Moreau, and C. William Sealey. (1992) “Fixed‐Rate Deposit Insurance and Risk‐
Shifting Behavior at Commercial Banks,” Journal of Banking and Finance, 16, 715‐42. Duan, Jin‐Chuan, and Jean‐Guy Simonato. (2002) “Maximum Likelihood Estimation of Deposit Insurance
Value with Interest‐Rate Risk,” Journal of Empirical Finance, 9, 109‐32. Eberlein, E. and D.B. Madan. (2010) Capital requirements, and taxpayer put option values for the major
US banks. Mimeo. Han, Chirok and Peter C.B. Phillips (2011) “First difference MLE and dynamic panel estimation”, Cowles Foundation Discussion Paper No. 1780. Heckman, James. (1976) “The Common Structure of Statistical Models of Truncation, Sample Selection
and Limited Dependent Variables and a Sample Estimator for Such Models,” Annals of Economic and Social Measurement, 5, 475‐92.
_________________. (1978) “Dummy Endogenous Variables in a Simultaneous Equation System,”
Econometrica, 46, 931‐59. Hovakimian, Armen, and Edward J. Kane. (2000) “Effectiveness of Capital Regulation at U.S. Commercial
Banks, 1985‐1994,” Journal of Finance, 55(March), 451‐469. Kane, Edward J., 1980. “Market Incompleteness and Divergences Between Forward and Future Interest
Rates,” Journal of Finance, 35, 221‐234. _________________. (1986). Appearance and reality for deposit institutions: The case for reform,
Journal of Banking and Finance, 175‐188. _________________. (2009) “Extracting Nontransparent Safety Net Subsidies by Strategically Expanding
and Contracting a Financial Institution’s Accounting Balance Sheet,” Journal of Financial Services Research, 36, 161‐68.
Marcus, Alan, and Israel Shaked. (1984) “The Valuation of FDIC Deposit Insurance Using Option‐Pricing
Estimates,” Journal of Money, Credit, and Banking, 16, 446‐460.
February 1, 2011
15
Merton, Robert C. (1977) “An Analytic Derivation of the Cost of Deposit Insurance and Loan Guarantees,” Journal of Banking and Finance, 1, 3‐11.
__________________. (1978). “on the Cost of Deposit Insurance When There Are Surveillance Costs,”
Journal of Business, 51, 439‐52. Molyneux, Philip, Klaus Schaeck, and Tim Mi Zhou. (2010) ‘Too‐Big‐to‐Fail’ and Its Impact on Safety Net
Subsidies and Systemic Risk, Bangor, Wales: Bangor University Working Paper. Ronn, Ehud, and A.R. Verma. (1986) “Pricing Risk‐Adjusted Deposit Insurance: An Option‐Based Model,”
Journal of Finance, 41, 871‐95.
February 1, 2011
16
TABLE 1 SAMPLE SIZE (NUMBER OF OBSERVATIONS)
Frequency: annual (2003-2008)
Austria 476
Belgium 627
Denmark 206
Finland 78
Luxembourg 426
Netherlands 203
Portugal 158
Sweden 263
Ireland 157
United Kingdom 864
Spain 531
France 1112
Italy 1236
Germany 2227
United States 2153
TOTAL 11117
February 1, 2011
17
TABLE 2
DEFINITIONS AND SOURCES FOR VARIABLES
Variable Definition Source B/V (%)
Leverage, measured as the ratio of the book value (B) of deposits and other debt to the market value of a bank’s assets
(V).
Bank-level data to compute this variable are obtained from the Bureau-Van Dijk
Bankscope database. IPP (%)
“Fair” insurance premium percentage, defined as the per-period flow of safety-net benefits that bank stockholders enjoy.
Bank-level data to compute this variable are obtained from the Bureau-Van Dijk
Bankscope database. V (%)
Volatility, defined as the standard deviation of the return on bank assets
Bank-level data to compute this variable are obtained from the Bureau-Van Dijk
Bankscope database. Size (log total
assets) (Eur mill) Size of the banks measured by total book value of assets.
Bank-level data to compute this variable are obtained from the Bureau-Van Dijk
Bankscope database. Corruption
perception index (10-CPI)
Transparency International’s Corruptions Perceptions Index (CPI) is an aggregate indicator that ranks countries in terms of
the degree to which corruption is perceived to exist among public officials and politicians. It is a composite index drawing
on corruption-related data by a variety of independent and reputable institutions. The main reason for using an aggregated
index of individual sources is that a combination of sources measuring the same phenomenon is more reliable than each
source taken separately. The CPI ranges 1 to 10. Higher values of the index show less corruption. In order to normalize the
values we have redefined the indicator as 10-CPI so that higher values show more corruption.
Transparency international (www.transparency.org)
Market volatility (VIX)
The VIX is calculated and disseminated in real-time by the Chicago Board Options Exchange. It is a weighted blend of
prices for a range of options on the S&P 500 index. On March 26, 2004, the first-ever trading in futures on the VIX Index
began on CBOE Futures Exchange (CFE). The formula uses a kernel-smoothed estimator that takes as inputs the current market prices for all out-of-the-money calls and puts for the front month and second month expirations.[1] The goal is to estimate the implied volatility of the S&P 500 index over the next 30 days. The VIX is the square root of the par variance
swap rate for a 30 day term initiated today. Note that the VIX is the volatility of a variance swap and not that of a volatility swap
(volatility being the square root of variance).
Chicago Board of Exchange (http://www.cboe.com/
micro/vix/introduction.aspx)
DFU Status A binary variable that takes on the value of unity for banks that
alternately either received open-bank assistance (DFUxp) or fell in the first decile of average 2003-2008 asset size for US and
European banks in the Bankscope database (DFUxa).
Deciles are calculated by the authors.
Identity of banks receiving equity injections is hand-
collected.
February 1, 2011
18
TABLE 3 MEAN LEVERAGE RATIO (B/V), MEAN FAIR PREMIUM (IPP), AND VOLATILITY OF RETURN ON
ASSETS (V): ALL BANKS, DFUxa and DFUxp BANKS IN EUROPE AND IN THE US
Country B/V (%) IPP (%) V (%) RV ML RV ML RV ML
ALL BANKS (FULL SAMPLE) 84.8 87.1 0.143 0.119 1.815 1.582 ALL BANKS IN EUROPE 85.3 86.0 0.153 0.134 1.988 1.727 ALL BANKS IN THE US 82.5 83.9 0.139 0.127 1.490 1.368
DFUxa BANKS (FULL SAMPLE) 86.9 89.8 0.167 0.145 1.593 1.597 DFUxp BANKS (FULL SAMPLE) 88.0 90.9 0.174 0.156 1.669 1.490
DFUxa BANKS IN EUROPE 88.1 90.0 0.179 0.164 1.696 1.487 DFUxp BANKS IN EUROPE 89.3 91.6 0.189 0.180 1.792 1.594 DFUxa BANKS IN THE US 80.5 82.2 0.127 0.116 1.396 1.284 DFUxp BANKS IN THE US 83.4 84.2 0.140 0.134 1.503 1.411
ALL BANKS IN EUROPE (PRE 2007) 86.7 88.0 0.157 0.163 2.134 2.166 ALL BANKS IN THE US (PRE 2007) 83.2 83.9 0.149 0.156 1.529 1.632
ALL BANKS IN EUROPE (2007-2008) 83.9 84.3 0.132 0.138 1.842 1.931 ALL BANKS IN THE US (2007-2008) 81.1 81.5 0.128 0.137 1.344 1.388
DFUxa BANKS IN EUROPE (PRE 2007) 90.4 92.6 0.198 0.185 1.591 1.403 DFUxa BANKS IN THE US (PRE 2007) 81.5 82.4 0.158 0.146 1.343 1.211
DFUxa BANKS IN EUROPE (2007-2008) 85.7 88.6 0.165 0.150 1.967 1.663 DFUxa BANKS IN THE US (2007-2008) 78.2 80.1 0.119 0.102 1.491 1.396
DFUxp BANKS IN EUROPE (PRE 2007) 92.3 93.4 0.215 0.220 1.635 1.523 DFUxp BANKS IN THE US (PRE 2007) 83.8 84.1 0.176 0.160 1.428 1.323
DFUxp BANKS IN EUROPE (2007-2008) 89.9 90.1 0.179 0.162 2.123 1.815 DFUxp BANKS IN THE US (2007-2008) 82.3 83.1 0.129 0.118 1.538 1.493
Mean difference tests: ALL BANKS IN EUROPE vs. ALL BANKS IN THE US 0.006 0.007 0.008 0.005 0.009 0.007 Mean difference tests: ALL BANKS vs. DFUxa BANKS (FULL SAMPLE) 0.012 0.011 0.008 0.004 0.005 0.006 Mean difference tests: ALL BANKS vs. DFUxp BANKS (FULL SAMPLE) 0.008 0.006 0.005 0.003 0.002 0.001
Mean difference tests: DFUxa vs. DFUxp BANKS (FULL SAMPLE) 0.005 0.004 0.003 0.001 0.001 0.001 Mean difference tests: ALL BANKS IN EUROPE vs. DFUxa BANKS IN
EUROPE 0.009 0.007 0.004 0.003 0.004 0.002
Mean difference tests: ALL BANKS IN THE US vs. DFUxa BANKS IN THE US 0.012 0.013 0.013 0.015 0.019 0.016 Mean difference tests: ALL BANKS IN EUROPE vs. DFUxp BANKS IN
EUROPE 0.007 0.005 0.0003 0.001 0.002 0.002
Mean difference tests: ALL BANKS IN THE US vs. DFUxa BANKS IN THE US 0.009 0.010 0.010 0.011 0.013 0.010 Mean difference tests: DFUxa BANKS IN EUROPE vs. DFUxp BANKS IN THE
US 0.002 0.003 0.005 0.004 0.007 0.006
Mean difference tests: DFUxp BANKS IN EUROPE vs. DFUxp BANKS IN THE US
0.001 0.001 0.003 0.002 0.004 0.004
Mean difference tests: ALL BANKS IN EUROPE (PRE 2007) vs. DFUxa BANKS IN EUROPE (PRE 2007)
0.005 0.003 0.004 0.002 0.004 0.003
Mean difference tests: ALL BANKS IN THE US (PRE 2007) vs. DFUxa BANKS IN THE US (PRE 2007)
0.012 0.014 0.005 0.004 0.006 0.004
Mean difference tests: ALL BANKS IN EUROPE (2007-2008) vs. DFUxa BANKS IN EUROPE (2007-2008)
0.004 0.003 0.005 0.003 0.002 0.004
Mean difference tests: ALL BANKS IN THE US (2007-2008) vs. DFUxa BANKS IN THE US (2007-2008)
0.008 0.011 0.005 0.007 0.004 0.005
Mean difference tests: DFUxa IN EUROPE (PRE 2007) VS. DFUxa IN EUROPE (2007-2008)
0.010 0.009 0.012 0.015 0.011 0.017
Mean difference tests: DFU IN THE US (PRE 2007) VS. DFUxa IN THE US (2007-2008)
0.008 0.007 0.010 0.005 0.013 0.010
Mean difference tests: ALL BANKS IN EUROPE (PRE 2007) vs. DFUxp BANKS IN EUROPE (PRE 2007)
0.001 0.001 0.001 0.001 0.002 0.002
Mean difference tests: ALL BANKS IN THE US (PRE 2007) vs. DFUxp BANKS IN THE US (PRE 2007)
0.008 0.009 0.003 0.002 0.004 0.003
Mean difference tests: ALL BANKS IN EUROPE (2007-2008) vs. DFUxp BANKS IN EUROPE (2007-2008)
0.002 0.002 0.003 0.001 0.001 0.002
Mean difference tests: ALL BANKS IN THE US (2007-2008) vs. DFUxp BANKS IN THE US (2007-2008)
0.002 0.001 0.002 0.004 0.002 0.003
Mean difference tests: DFUxp IN EUROPE (PRE 2007) VS. DFUxp IN EUROPE (2007-2008)
0.008 0.006 0.010 0.011 0.008 0.013
Mean difference tests: DFU IN THE US (PRE 2007) VS. DFUxp IN THE US (2007-2008)
0.006 0.004 0.006 0.004 0.010 0.007
All estimated parameters are significant at the 1% level The test statistics report the p–value of a one–tailed t–test of the hypothesis that the means are equal for the indicated groups.
February 1, 2011
19
TABLE 4
SINGLE-EQUATION ESTIMATES OF THE EFFECTIVENESS OF SAFETY-NET CONTROL: ALL BANKS AND DFUxa BANKS IN EUROPE AND IN THE US
Fixed-effects panel regressions relating changes in a bank’s leverage, B/V, and changes in its fair insurance premium percentage, IPP, to the changes in volatility of its assets, V. B is the face value of bank’s debt, including
deposits. V is the market value of bank assets. The errors are clustered at the firm level
European sample (B/V) IPP RV ML RV ML
V -0.002** (-26.14)
-0.004** (-34.17)
0.007** (19.83)
0.008** (25.16)
Size (log total assets) 0.013** (14.31)
0.016** (17.90)
-0.015** (-14.51)
-0.011** (16.31)
V X DFUxa banks Europe -0.020** (-6.53)
-0.025** (-8.83)
0.019** (6.50)
0.020** (7.28)
Size X DFUxa banks Europe 0.003 (1.23)
0.001 (1.01)
0.003 (1.23)
0.003 (1.23)
Corruption perception index (10-CPI)
0.008** (3.29)
0.011** (4.88)
0.016** (6.04)
0.008** (3.29)
Market volatility (VIX) -0.001* (1.93)
-0.001* (2.16)
0.012 (0.27)
0.018 (0.14)
Observations 8,964 8,964 8,964 8,964 Number of banks 1,494 1,494 1,494 1,494
R2 0.517 0.604 0.685 0.643
US sample RV ML RV ML
V -0.006** (-18.07)
-0.007** (-31.20)
0.009** (18.51)
0.011** (25.14)
Size (log total assets) 0.029** (14.13)
0.024** (17.53)
-0.016** (-11.15)
-0.014** (22.23)
V X DFUxa banks US -0.038** (-5.57)
-0.032** (-8.92)
0.024** (3.63)
0.029** (3.97)
Size X DFUxa banks US 0.002 (1.12)
0.004 (1.25)
0.007 (0.44)
0.003 (0.78)
Corruption perception index (10-CPI)
0.004 (1.18)
0.007 (0.96)
0.010 (0.85)
0.006 (0.72)
Market volatility (VIX) -0.003** (2.85)
-0.004** (3.49)
0.010 (0.68)
0.012 (0.19)
Observations 2,153 2,153 2,153 2,153 Number of banks 358 358 358 358
R2 0.693 0.618 0.688 0.715
Test of the differences between the European and the US sample (p-value) V 0.020 0.018 0.013 0.014
Size (log total assets) 0.004 0.009 0.198 0.032 V X DFUxa banks US 0.003 0.036 0.013 0.009 Size X DFUxa banks US 0.002 0.011 0.011 0.396
Corruption perception index (10-CPI)
0.023 0.028 0.021 0.024
Market volatility (VIX) 0.059 0.053 0.061 0.036 * Statistically significant at 5% level ** Statistically significant at 1% level
February 1, 2011
20
TABLE 5A
SINGLE-EQUATION ESTIMATES OF THE EFFECTIVENESS OF SAFETY-NET CONTROL: ALL BANKS AND DFUxp BANKS (BENEFITING FROM STATE AID) IN EUROPE AND IN
THE US Fixed-effects panel regressions relating changes in a bank’s leverage, B/V, and changes in its fair insurance premium percentage, IPP, to changes in the volatility of its assets, V. B is the face value of bank’s debt, including deposits. V is the market value of bank assets. The errors
are clustered at the firm level
European sample (B/V) IPP RV ML RV ML
V -0.003** (-18.31)
-0.005** (-22.51)
0.006** (14.02)
0.007** (33.08)
Size (log total assets) 0.011** (12.24)
0.014** (18.88)
-0.013** (-17.29)
-0.010** (14.25)
V X DFUxp banks in Europe -0.009** (-7.12)
-0.012** (-7.31)
0.027** (8.15)
0.029** (6.10)
Corruption perception index (10-CPI)
0.010** (2.98)
0.011** (4.88)
0.016** (6.04)
0.008** (3.29)
Market volatility (VIX) -0.002* (2.20)
-0.007** (2.96)
0.013 (0.08)
0.011 (0.19)
Observations 8,964 8,964 8,964 8,964 Number of banks 1,494 1,494 1,494 1,494
R2 0.616 0.594 0.702 0.625
US sample RV ML RV ML
V -0.006** (-17.12)
-0.008** (-28.68)
0.010** (17.27)
0.013** (22.65)
Size (log total assets) 0.025** (16.77)
0.019** (14.31)
-0.018** (-12.72)
-0.017** (25.90)
V X DFUxp banks in the US -0.022** (-6.19)
-0.028** (-6.84)
0.033* (2.14)
0.035** (4.42)
Corruption perception index (10-CPI)
0.003 (0.82)
0.005 (0.48)
0.014 (1.12)
0.010 (0.95)
Market volatility (VIX) -0.006** (3.48)
-0.005** (3.89)
0.014 (0.71)
0.011 (0.28)
Observations 2,153 2,153 2,153 2,153 Number of banks 358 358 358 358
R2 0.685 0.624 0.603 0.745 Test of the differences between the European and the US sample (p-value)
V 0.016 0.014 0.016 0.014 Size (log total assets) 0.004 0.006 0.004 0.006
V X DFUxp banks in Europe 0.003 0.002 0.003 0.002 Corruption perception index
(10-CPI) 0.002 0.005 0.002 0.005
Market volatility (VIX) 0.005 0.016 0.005 0.016 * Statistically significant at 5% level ** Statistically significant at 1% level
February 1, 2011
21
TABLE 5B SINGLE-EQUATION ESTIMATES OF THE EFFECTIVENESS OF SAFETY-NET CONTROL: ALL BANKS AND
DFUxp BANKS IN EUROPE AND THE US Fixed-effects panel regressions relating changes in a bank’s leverage, B/V, and changes in its fair insurance premium percentage, IPP, to changes in the volatility of its assets, V. B is the face value of bank’s debt, including deposits. V is the market value of
bank assets. The errors are clustered at the firm level
European sample (B/V) IPP RV ML RV ML
V -0.004** (-14.26)
-0.006** (-21.05)
0.005** (13.04)
0.007** (28.14)
Lambda (Mills ratio) -0.058* (1.99)
-0.081** (3.93)
-0.028** (10.13)
-0.034** (7.82)
Size (log total assets) 0.010** (11.51)
0.016** (17.23)
-0.011** (-17.50)
-0.013** (13.85)
V X DFUxp banks in Europe -0.009** (-6.14)
-0.013** (-7.18)
0.029** (8.96)
0.025** (5.08)
Corruption perception index (10-CPI)
0.011** (2.08)
0.014** (5.15)
0.013** (6.17)
0.004** (3.22)
Market volatility (VIX) -0.002* (2.14)
-0.006** (3.17)
0.012 (0.19)
0.014 (0.11)
Observations 8,964 8,964 8,964 8,964 Number of banks 1,494 1,494 1,494 1,494
R2 0.649 0.629 0.718 0.632 FIXED‐EFFECTS PROBIT SELECTION MODELS FOR ZERO‐ONE BINARY VARIABLES DISTINGUISHING DFU BANKS BENEFITING FROM STATE AID (1)
FROM THE REST OF DFU BANKS (0) V 0.963**
(12.39) 0.815** (7.05)
0.963** (12.39)
0.815** (7.05)
Size (log total assets) 0.013 (1.16)
0.004 (0.96)
0.013 (1.16)
0.004 (0.96)
Corruption perception index (10-CPI)
0.823** (6.28)
0.916** (8.62)
0.823** (6.28)
0.916** (8.62)
Observations 826 826 826 826 Number of DFUxa banks 137 137 137 137 Number of DFUxp banks 43 43 43 43
Log-likelihood -626.3 -458.5 -626.3 -458.5 Fraction of correct predictions 88.5 90.4 88.5 90.4
US sample V -0.007**
(-14.06) -0.24.06) 0.011**
(13.08) 0.012** (21.04)
Lambda (Mills ratio) -0.094** (4.41)
-0.078** (5.13)
-0.028** (6.40)
-0.034** (6.21)
Size (log total assets) 0.028** (15.93)
0.020** (11.10)
-0.016** (-12.13)
-0.013** (23.03)
V X DFUxp banks in the US -0.021** (-7.05)
-0.031** (-7.13)
0.034* (2.10)
0.030** (5.06)
Corruption perception index (10-CPI)
0.005 (0.88)
0.006 (0.51)
0.013 (1.08)
0.009 (0.72)
Market volatility (VIX) -0.006** (3.20)
-0.007** (4.13)
0.014 (0.62)
0.012 (0.33)
Observations 2,153 2,153 2,153 2,153 Number of banks 358 358 358 358
R2 0.690 0.645 0.615 0.758 FIXED‐EFFECTS PROBIT SELECTION MODELS FOR ZERO‐ONE BINARY VARIABLES DISTINGUISHING DFU BANKS BENEFITING FROM STATE AID (1)
FROM THE REST OF DFU BANKS (0) V 0.703**
(18.05) 0.626** (12.35)
0.703** (18.05)
0.626** (12.35)
Size (log total assets) 1.624** (6.51)
1.498** (7.18)
1.624** (6.51)
1.498** (7.18)
Corruption perception index (10-CPI)
0.621 (0.44)
0.521 (0.76)
0.621 (0.44)
0.521 (0.76)
Observations 203 203 203 203 Number of DFUxa banks 33 33 33 33 Number of DFUxp banks 22 22 22 22
Log-likelihood -484.0 -507.2 -484.0 -507.2 Fraction of correct predictions 89.9 88.5 89.9 88.5
February 1, 2011
22
Test of the differences between the European and the US sample (p-value) V 0.015 0.016 0.021 0.012
Size (log total assets) 0.003 0.005 0.007 0.006 V X DFUxp banks 0.005 0.002 0.011 0.004
Corruption perception index (10-CPI)
0.002 0.004 0.041 0.045
Market volatility (VIX) 0.007 0.015 0.596 0.624 * Statistically significant at 5% level ** Statistically significant at 1% level
February 1, 2011
23
TABLE 6A
SINGLE-EQUATION ESTIMATES OF THE EFFECTIVENESS OF SAFETY-NET CONTROL: PRE-CRISIS ENVIRONMENT (2003-2006)
Fixed-effects panel regressions relating changes in a bank’s leverage, B/V, and changes in its fair insurance premium percentage, IPP, to changes in the volatility of its assets, V. B is the face value of bank’s debt, including deposits. V is the market value of bank assets. The errors
are clustered at the firm level
European sample (B/V) IPP RV ML RV ML
V -0.003** (-29.47)
-0.005** (-33.42)
0.005** (14.24)
0.007** (22.73)
Size (log total assets) 0.015** (12.10)
0.018** (19.81)
-0.018** (-15.46)
-0.014** (13.78)
V X DFUxa banks Europe -0.029** (-7.77)
-0.030** (-6.76)
0.024** (4.72)
0.026** (4.84)
Corruption perception index (10-CPI)
0.011** (5.18)
0.005** (4.93)
0.011** (5.91)
0.007** (4.58)
Market volatility (VIX) -0.001* (2.19)
-0.001** (2.84)
0.018 (0.52)
0.025 (0.27)
Observations 6,156 6,156 6,156 6,156 Number of banks 1,539 1,539 1,539 1,539
R2 0.517 0.562 0.597 0.534
US sample RV ML RV ML
V -0.005** (-16.35)
-0.009** (-24.15)
0.011** (12.16)
0.013** (16.31)
Size (log total assets) 0.037** (15.56)
0.032** (16.74)
-0.013** (-14.20)
-0.011** (18.26)
V X DFUxa banks US -0.041* (-2.21)
-0.035** (-2.19)
0.032** (5.84)
0.034** (3.13)
Corruption perception index (10-CPI)
0.002 (0.77)
0.003 (0.53)
0.012 (0.97)
0.008 (0.68)
Market volatility (VIX) -0.004** (3.99)
-0.005** (5.18)
0.008 (0.76)
0.018 (0.21)
Observations 1,398 1,398 1,398 1,398 Number of banks 349 349 349 349
R2 0.584 0.494 0.652 0.626 Test of the differences between the European and the US sample (p-value)
V 0.014 0.011 0.015 0.012 Size (log total assets) 0.002 0.003 0.006 0.026
V X DFUxa banks Europe 0.003 0.031 0.078 0.003 Corruption perception index
(10-CPI) 0.006 0.031 0.362 0.408
Market volatility (VIX) 0.023 0.013 0.014 0.008 * Statistically significant at 5% level ** Statistically significant at 1% level
February 1, 2011
24
TABLE 6B SINGLE-EQUATION ESTIMATES OF THE EFFECTIVENESS OF SAFETY-NET CONTROL:
CRISIS ENVIRONMENT (2007-2008) Fixed-effects panel regressions relating changes in a bank’s leverage, B/V, and changes in its fair insurance premium percentage, IPP, to changes in the volatility of its assets, V. B is the face value of bank’s debt, including deposits. V is the market value of bank assets. The errors
are clustered at the firm level.
European sample (B/V) IPP RV ML RV ML
V -0.005** (-20.30)
-0.006** (-31.51)
0.009** (12.51)
0.011** (16.83)
Size (log total assets) 0.015** (16.71)
0.018** (13.95)
-0.014** (-13.38)
-0.009** (10.14)
V X DFUxa banks Europe -0.025** (-4.96)
-0.020* (-2.17)
0.033** (6.36)
0.029** (4.57)
Corruption perception index (10-CPI)
0.015** (6.85)
0.009** (4.08)
0.014** (3.63)
0.012** (3.90)
Market volatility (VIX) -0.004** (3.52)
-0.003** (2.92)
0.025 (0.63)
0.032 (0.44)
Observations 2,808 2,808 2,808 2,808 Number of banks 1,404 1,404 1,404 1,404
R2 0.501 0.495 0.542 0.512
US sample RV ML RV ML
V -0.008** (-12.64)
-0.009** (-31.20)
0.012** (14.24)
0.015** (20.44)
Size (log total assets) 0.028** (12.41)
0.027** (11.14)
-0.010** (-12.35)
-0.017** (16.51)
V X DFUxa banks US -0.024** (-4.50)
-0.023** (8.92)
0.039** (4.27)
0.044** (5.23)
Corruption perception index (10-CPI)
0.001 (0.63)
0.002 (0.32)
0.010 (0.59)
0.006 (0.70)
Market volatility (VIX) -0.005** (5.02)
-0.004** (4.28)
0.006 (0.44)
0.012 (0.58)
Observations 755 755 755 755 Number of banks 377 377 377 377
R2 0.602 0.528 0.538 0.586
Test of the differences in V between the European and the US sample (p-value) V 0.012 0.015 0.013 0.010
Size (log total assets) 0.002 0.003 0.010 0.003 V X DFUxa banks Europe 0.126 0.037 0.004 0.002 Corruption perception index
(10-CPI) 0.002 0.012 0.006 0.007
Market volatility (VIX) 0.586 0.489 0.008 0.004 * Statistically significant at 5% level ** Statistically significant at 1% level
February 1, 2011
25
TABLE 7 TESTS OF DIFFERENCES BETWEEN PRECRISIS (2003-2006) AND CRISIS YEARS (2007-
2008) FOR THE US AND EUROPE SEPARATELY The table shows p-values of covariance tests for coefficient differences as well as the F-test of the overall differences
between the sub-samples
European sample (B/V) IPP RV ML RV ML
V 0.023 0.042 0.007 0.012
Size (log total assets) 0.653 0.728 0.046 0.007
V X DFUxa banks Europe 0.124 0.005 0.006 0.138
Corruption perception index (10-CPI)
0.014 0.011 0.088 0.005
Market volatility (VIX) 0.009 0.016 0.008 0.018
Overall coefficients F-test 0.018 0.013 0.011 0.016
US sample (B/V) IPP RV ML RV ML
V 0.046 0.196 0.140 0.051
Size (log total assets) 0.007 0.006 0.013 0.004
V X DFUxa banks US 0.003 0.008 0.053 0.009
Corruption perception index (10-CPI)
0.963 0.694 0.121 0.160
Market volatility (VIX) 0.864 0.658 0.134 0.079
Overall coefficients F-test 0.019 0.034 0.038 0.030
February 1, 2011
26
TABLE 8A SINGLE-EQUATION ESTIMATES OF THE EFFECTIVENESS OF SAFETY-NET CONTROL: ALL BANKS AND
DFUxp BANKS IN EUROPE AND IN THE US PRE-CRISIS ENVIRONMENT (2003-2006)
Fixed-effects panel regressions relating changes in a bank’s leverage, B/V, and changes in its fair insurance premium percentage, IPP, to changes in the riskiness of its assets, V. B is the face value of bank’s debt, including deposits. V is the market value of
bank assets. The errors are clustered at the firm level
European sample (B/V) IPP RV ML RV ML
V -0.003** (-15.64)
-0.004** (-12.59)
0.005** (11.90)
0.004** (21.03)
Lambda (Mills ratio) -0.061* (1.95)
-0.073** (3.26)
-0.024** (10.03)
-0.038** (7.30)
Size (log total assets) 0.008** (8.41)
0.009** (11.57)
-0.014** (-13.82)
-0.007** (10.13)
V X DFUxp banks in Europe -0.006** (-4.14)
-0.011** (-7.23)
0.021** (5.63)
0.025** (4.52)
Corruption perception index (10-CPI)
0.005** (4.23)
0.008** (3.48)
0.011** (4.52)
0.014** (5.27)
Market volatility (VIX) -0.004** (3.94)
-0.005** (2.86)
0.007 (0.28)
0.005 (0.25)
Observations 6,156 6,156 6,156 6,156 Number of banks 1,539 1,539 1,539 1,539
R2 0.492 0.536 0.559 0.580 FIXED‐EFFECTS PROBIT SELECTION MODELS FOR ZERO‐ONE BINARY VARIABLES DISTINGUISHING DFU BANKS BENEFITING FROM STATE AID (1)
FROM THE REST OF DFU BANKS (0) V 0.826**
(8.13) 0.803** (4.28)
0.826** (8.13)
0.803** (4.28)
Size (log total assets) 0.010 (1.03)
0.002 (0.31)
0.010 (1.03)
0.002 (0.31)
Corruption perception index (10-CPI)
0.426** (4.13)
0.531** (4.26)
0.426** (4.13)
0.531** (4.26)
Observations 534 534 534 534 Number of DFUxa banks 133 133 133 133 Number of DFUxp banks 43 43 43 43
Log-likelihood -412.8 -469.1 -412.8 -469.1 Fraction of correct predictions 87.4 88.5 87.4 88.5
US sample V -0.004**
(-11.77) -0.21.07) 0.006**
(10.13) 0.011** (14.93)
Lambda (Mills ratio) -0.074** (4.14)
-0.060** (4.92)
-0.024** (5.31)
-0.031** (5.55)
Size (log total assets) 0.013** (12.65)
0.014* (2.01)
-0.014* (-2.38)
-0.013** (19.04)
V X DFUxp banks in the US -0.016** (-8.20)
-0.020** (-5.40)
0.030** (2.73)
0.034** (3.81)
Corruption perception index (10-CPI)
0.004 (0.93)
0.003 (0.53)
0.010 (1.52)
0.006 (1.20)
Market volatility (VIX) -0.007** (3.07)
-0.005** (4.29)
0.007 (0.93)
0.012 (0.32)
Observations 1398 1398 1398 1398 R2 0.586 0.469 0.626 0.590
FIXED‐EFFECTS PROBIT SELECTION MODELS FOR ZERO‐ONE BINARY VARIABLES DISTINGUISHING DFU BANKS BENEFITING FROM STATE AID (1) FROM THE REST OF DFU BANKS (0)
V 0.423** (12.31)
0.415** (7.80)
0.423** (12.31)
0.415** (7.80)
Size (log total assets) 0.840** (8.13)
0.902** (4.03)
0.840** (8.13)
0.902** (4.03)
Corruption perception index (10-CPI)
0.403 (0.32)
0.491 (0.86)
0.403 (0.32)
0.491 (0.86)
Observations 128 128 128 128 Number of DFUxa banks 32 32 32 32 Number of DFUxp banks 22 22 22 22
Log-likelihood -412.7 -477.7 -412.7 -477.7 Fraction of correct predictions 86.8 88.2 86.8 88.2
Test of the differences between the European and the US sample (p-value) V 0.025 0.020 0.019 0.016
Size (log total assets) 0.002 0.012 0.006 0.008 V X DFUxp banks in the US 0.005 0.007 0.026 0.003
Corruption perception index (10-CPI) 0.117 0.016 0.059 0.036 Market volatility (VIX) 0.029 0.014 0.631 0.017
February 1, 2011
27
Test of the differences between the selection equations for the European and the US sample (p-value) V 0.002 0.001 0.003 0.004
Size (log total assets) 0.001 0.002 0.001 0.001 Corruption perception index (10-CPI) 0.013 0.019 0.003 0.001
* Statistically significant at 5% level ** Statistically significant at 1% level
February 1, 2011
28
TABLE 8B SINGLE-EQUATION ESTIMATES OF THE EFFECTIVENESS OF SAFETY-NET CONTROL: ALL BANKS AND
DFUxp BANKS IN EUROPE AND IN THE US DURING CRISIS YEARS (2007-2008)
Fixed-effects panel regressions relating changes in a bank’s leverage, B/V, and changes in its fair insurance premium percentage, IPP, to changes in the riskiness of its assets, V. B is the face value of bank’s debt, including deposits. V is the market value of
bank assets. The errors are clustered at the firm level
European sample (B/V) IPP RV ML RV ML
V -0.004** (-12.03)
-0.006** (-14.28)
0.006** (10.37)
0.007** (24.28)
Lambda (Mills ratio) -0.052* (1.84)
-0.062** (3.81)
-0.014** (4.53)
-0.035** (7.60)
Size (log total assets) 0.006** (8.21)
0.015** (9.51)
-0.011** (-11.17)
-0.012** (13.68)
V X DFUxp banks in Europe -0.010** (-7.35)
-0.018* (-2.13)
0.033** (7.08)
0.036** (5.27)
Corruption perception index (10-CPI)
0.007** (6.13)
0.007** (4.27)
0.012** (3.29)
0.018** (3.70)
Market volatility (VIX) -0.005** (3.23)
-0.004** (4.02)
0.006 (0.25)
0.008 (0.56)
Observations 2,808 2,808 2,808 2,808
Number of banks 1,404 1,404 1,404 1,404 R2 0.475 0.443 0.518 0.468
FIXED‐EFFECTS PROBIT SELECTION MODELS FOR ZERO‐ONE BINARY VARIABLES DISTINGUISHING DFU BANKS BENEFITING FROM STATE AID (1) FROM THE REST OF DFU BANKS (0)
V 0.285** (6.01)
0.327** (5.32)
0.285** (6.01)
0.327** (5.32)
Size (log total assets) 0.004 (1.23)
0.005 (0.90)
0.004 (1.23)
0.005 (0.90)
Corruption perception index (10-CPI)
0.415* (1.97)
0.885** (6.17)
0.415* (1.97)
0.885** (6.17)
Observations 292 292 292 292 Number of DFUxa banks 146 146 146 146 Number of DFUxp banks 43 43 43 43
Log-likelihood -348.3 -435.3 -348.3 -435.3 Fraction of correct predictions 85.2 87.9 85.2 87.9
US sample RV ML RV ML
V -0.007** (-12.16)
-0.009** (-12.57)
0.015** (11.16)
0.013** (12.83)
Lambda (Mills ratio) -0.046** (6.13)
-0.079** (5.28)
-0.024** (6.32)
-0.036** (5.04)
Size (log total assets) 0.030* (2.16)
0.020** (4.86)
-0.019* (-4.27)
-0.018* (-1.63)
V X DFUxp banks in the US -0.033** (-7.95)
-0.008** (-3.53)
0.034* (2.08)
0.042** (4.20)
Corruption perception index (10-CPI)
0.007 (0.63)
0.006 (0.63)
0.015 (1.03)
0.010 (0.90)
Market volatility (VIX) -0.010** (3.38)
-0.012** (4.02)
0.006 (0.20)
0.008 (0.24)
Observations 755 755 755 755 Number of banks 377 377 377 377
R2 0.572 0.477 0.506 0.547 FIXED‐EFFECTS PROBIT SELECTION MODELS FOR ZERO‐ONE BINARY VARIABLES DISTINGUISHING DFU BANKS BENEFITING FROM STATE AID (1)
FROM THE REST OF DFU BANKS (0) V 0.982**
(4.01) 0.785** (3.68)
0.982** (4.01)
0.785** (3.68)
Size (log total assets) 1.626** (5.13)
1.494** (5.92)
1.626** (5.13)
1.494** (5.92)
Corruption perception index (10-CPI)
0.574 (0.42)
0.546 (0.58)
0.574 (0.42)
0.546 (0.58)
Observations 75 75 75 75 Number of DFUxa banks 37 37 37 37 Number of DFUxp banks 22 22 22 22
Log-likelihood -390.7 -349.7 -390.7 -349.7 Fraction of correct predictions 86.3 88.7 86.3 88.7
February 1, 2011
29
Test of the differences between the European and the US sample (p-value) V 0.020 0.018 0.015 0.013
Size (log total assets) 0.003 0.006 0.004 0.007 V X DFUxp banks in the US 0.005 0.003 0.044 0.038
Corruption perception index (10-CPI) 0.705 0.020 0.029 0.027 Market volatility (VIX) 0.013 0.014 0.657 0.266
Test of the differences between the selection equations for the European and the US sample (p-value) V 0.001 0.001 0.001 0.001
Size (log total assets) 0.001 0.001 0.001 0.001 Corruption perception index (10-CPI) 0.023 0.006 0.002 0.002
* Statistically significant at 5% level ** Statistically significant at 1% level
February 1, 2011
30
TABLE 9 TESTS OF THE DIFFERENCES BETWEEN THE PRECRISIS (2003-2006) AND CRISIS
YEARS (2007-2008): DFUxp BANKS IN EUROPE AND THE US The table show the p-values of the tests for coefficient differences as well as the F-test of the overall differences
between the subsamples
European sample (B/V) IPP RV ML RV ML
V 0.006 0.012 0.006 0.003
Size (log total assets) 0.011 0.004 0.107 0.052
V X DFUxp banks in Europe 0.008 0.009 0.004 0.006
Corruption perception index (10-CPI)
0.019 0.294 0.013 0.011
Market volatility (VIX) 0.048 0.031 0.002 0.003
Overall coefficients F-test 0.010 0.013 0.008 0.009
US sample (B/V) IPP RV ML RV ML
V 0.005 0.011 0.004 0.003
Size (log total assets) 0.003 0.003 0.003 0.004
V X DFUxp banks in the US 0.004 0.005 0.005 0.128
Corruption perception index (10-CPI)
0.043 0.238 0.013 0.029
Market volatility (VIX) 0.031 0.011 0.002 0.003
Overall coefficients F-test 0.008 0.010 0.006 0.008
February 1, 2011
31
TABLE 10A SINGLE-EQUATION ESTIMATES OF THE EFFECTIVENESS OF SAFETY-NET CONTROL: PORTUGAL AND
IRELAND Fixed-effects panel regressions relating changes in a bank’s leverage, B/V, and changes in its fair insurance premium percentage, IPP, to changes in the riskiness of its assets, V. B is the face value of bank’s debt, including deposits. V is the market value of
bank assets. The errors are clustered at the firm level
Portugal (B/V) IPP RV ML RV ML
V -0.003** (-8.64)
-0.004** (-22.16)
0.011** (10.51)
0.010** (9.87)
Lambda (Mills ratio) -0.043** (2.98)
-0.071** (5.53)
-0.019** (6.18)
-0.024** (5.96)
Size (log total assets) 0.006** (11.13)
0.008** (13.82)
-0.014** (-9.32)
-0.013** (12.63)
V X DFUxp banks in Portugal -0.012** (-3.98)
-0.015** (-5.83)
0.038** (2.94)
0.030** (2.61)
Corruption perception index (10-CPI)
0.010** (3.31)
0.009** (2.31)
0.014** (5.02)
0.011** (3.36)
Market volatility (VIX) -0.004 (1.03)
-0.003 (1.27)
0.005 (0.33)
0.003 (0.28)
Observations 158 158 158 158 Number of banks 26 26 26 26
R2 0.403 0.460 0.484 0.520 FIXED‐EFFECTS PROBIT SELECTION MODELS FOR ZERO‐ONE BINARY VARIABLES DISTINGUISHING DFU BANKS BENEFITING FROM STATE AID (1)
FROM THE REST OF DFU BANKS (0) V 0.423**
(5.18) 0.432** (6.15)
0.423** (5.18)
0.432** (6.15)
Size (log total assets) 0.026 (1.63)
0.012 (0.94)
0.026 (1.63)
0.012 (0.94)
Corruption perception index (10-CPI)
0.661* (2.23)
0.891** (6.02)
0.661* (2.23)
0.891** (6.02)
Observations 24 24 24 24 Number of DFUxa banks 4 4 4 4 Number of DFUxp banks 2 2 2 2
Log-likelihood -326.7 -460.3 -326.7 -460.3 Fraction of correct predictions 86.4 88.4 86.4 88.4
Ireland RV ML RV ML
V -0.002** (-11.01)
-0.002** (-16.50)
0.015** (13.08)
0.018** (14.82)
Lambda (Mills ratio) -0.039* (2.63)
-0.050** (4.09)
-0.028** (6.08)
-0.042** (9.02)
Size (log total assets) 0.023** (12.60)
0.013** (19.42)
-0.018** (-10.09)
-0.020** (14.37)
V X DFUxp banks in Ireland -0.018** (-6.54)
-0.019** (-3.88)
0.049** (3.62)
0.064** (4.03)
Corruption perception index (10-CPI)
0.084** (5.23)
0.041** (5.03)
0.018** (5.21)
0.021** (5.52)
Market volatility (VIX) -0.006 (1.32)
-0.005 (1.62)
0.016 (0.40)
0.019 (0.28)
Observations 157 157 157 157 Number of banks 25 25 25 25
R2 0.447 0.416 0.593 0.496 FIXED‐EFFECTS PROBIT SELECTION MODELS FOR ZERO‐ONE BINARY VARIABLES DISTINGUISHING DFU BANKS BENEFITING FROM STATE AID (1) FROM THE REST
OF DFU BANKS (0) V 0.225**
(4.36) 0.212** (5.28)
0.225** (4.36)
0.212** (5.28)
Size (log total assets) 0.096* (2.13)
0.077* (4.82)
0.096* (2.13)
0.077* (4.82)
Corruption perception index (10-CPI)
0.686** (2.91)
0.719** (4.64)
0.686** (2.91)
0.719** (4.64)
Observations 24 24 24 24 Number of DFUxa banks 4 4 4 4 Number of DFUxp banks 3 3 3 3
Log-likelihood -3263 -412.0 -3263 -412.0 Fraction of correct predictions 83.7 85.3 83.7 85.3
* Statistically significant at 5% level ** Statistically significant at 1% level
February 1, 2011
32
TABLE 10B
SINGLE-EQUATION ESTIMATES OF THE EFFECTIVENESS OF SAFETY-NET CONTROL: SPAIN AND ITALY
Fixed-effects panel regressions relating changes in a bank’s leverage, B/V, and changes in its fair insurance premium percentage, IPP, to changes in the riskiness of its assets, V. B is the face value of bank’s debt, including deposits. V is the market value of
bank assets. The errors are clustered at the firm level
Spain (B/V) IPP RV ML RV ML
V -0.004** (-15.04)
-0.005** (-25.18)
0.006** (16.12)
0.008** (16.27)
Lambda (Mills ratio) -0.021** (4.88)
-0.025** (6.58)
-0.016** (3.31)
-0.014** (4.19)
Size (log total assets) 0.037 (0.85)
0.027 (0.31)
0.012 (0.63)
0.010 (0.40)
V X DFUxp banks in Spain -0.018** (-6.03)
-0.020** (-5.13)
0.019** (3.34)
0.023** (4.16)
Corruption perception index (10-CPI)
0.004** (3.18)
0.003** (2.58)
0.007** (5.01)
0.008** (3.14)
Market volatility (VIX) -0.002 (0.73)
-0.001 (0.34)
0.005 (0.20)
0.004 (0.33)
Observations 531 531 531 531 Number of banks 86 86 86 86
R2 0.503 0.550 0.519 0.523 FIXED‐EFFECTS PROBIT SELECTION MODELS FOR ZERO‐ONE BINARY VARIABLES DISTINGUISHING DFU BANKS BENEFITING FROM STATE AID (1)
FROM THE REST OF DFU BANKS (0) V 0.131**
(4.56) 0.131** (5.01)
0.131** (4.56)
0.131** (5.01)
Size (log total assets) 0.008 (0.55)
0.009 (0.65)
0.008 (0.55)
0.009 (0.65)
Corruption perception index (10-CPI)
0.285** (2.76)
0.826** (6.04)
0.285** (2.76)
0.826** (6.04)
Observations 52 52 52 52 Number of DFUxa banks 8 8 8 8 Number of DFUxp banks 4 4 4 4
Log-likelihood -318.5 -401.7 -318.5 -401.7 Fraction of correct predictions 83.7 84.2 83.7 84.2
Italy V -0.007**
(-10.13) -0.008** (-15.06)
0.008** (16.67)
0.006** (12.34)
Lambda (Mills ratio) -0.059* (2.31)
-0.063* (2.13)
-0.052** (4.83)
-0.060** (7.15)
Size (log total assets) 0.032** (13.84)
0.038** (14.13)
-0.029** (-7.15)
-0.010** (14.32)
V X DFUxp banks in Italy -0.016** (-7.50)
-0.023** (-8.31)
0.026** (4.77)
0.020** (5.32)
Corruption perception index (10-CPI)
0.008** (4.83)
0.006** (4.94)
0.005** (3.34)
0.006** (4.05)
Market volatility (VIX) -0.001 (0.30)
-0.002 (0.35)
0.003 (0.62)
0.004 (0.30)
Observations 1236 1236 1236 1236 Number of DFU banks 206 206 206 206
R2 0.576 0.593 0.580 0.613 FIXED‐EFFECTS PROBIT SELECTION MODELS FOR ZERO‐ONE BINARY VARIABLES DISTINGUISHING DFU BANKS BENEFITING FROM STATE AID (1)
FROM THE REST OF DFU BANKS (0) V 0.721**
(7.58) 0.850** (4.16)
0.721** (7.58)
0.850** (4.16)
Size (log total assets) 0.013* (1.96)
0.010** (3.05)
0.013* (1.96)
0.010** (3.05)
Corruption perception index (10-CPI)
0.421** (3.75)
0.478** (4.01)
0.421** (3.75)
0.478** (4.01)
Observations 120 120 120 120 Number of DFUxa banks 20 20 20 20 Number of DFUxp banks 6 6 6 6
Log-likelihood -360.5 -390.0 -360.5 -390.0 Fraction of correct predictions 85.0 86.5 85.0 86.5
February 1, 2011
33
TABLE 10C TESTS OF THE DIFFERENCES BETWEEN PORTUGAL, IRELAND,
SPAIN AND ITALY DFUxp BANKS p-values in parentheses
Test of the differences between Portugal and Ireland V 0.057 0.046 0.026 0.022
Size (log total assets) 0.016 0.010 0.005 0.004 V X DFUxp banks in Spain 0.002 0.006 0.030 0.023
Corruption perception index (10-CPI) 0.006 0.002 0.028 0.019 Market volatility (VIX) 0.014 0.030 0.016 0.008
Test of the differences in the selection equation between Portugal and Ireland V 0.004 0.003 0.003 0.002
Size (log total assets) 0.002 0.001 0.002 0.010 Corruption perception index (10-CPI) 0.006 0.023 0.003 0.003
Test of the differences between Portugal and Spain V 0.326 0.286 0.015 0.039
Lambda (Mills ratio) 0.003 0.004 0.042 0.026 Size (log total assets) 0.002 0.002 0.123 0.086
V X DFUxp banks in Italy 0.024 0.027 0.006 0.007 Corruption perception index (10-CPI) 0.028 0.031 0.001 0.001
Market volatility (VIX) 0.263 0.385 0.698 0.582 Test of the differences in the selection equation between Portugal and Spain
V 0.002 0.002 0.001 0.003 Size (log total assets) 0.003 0.004 0.638 0.125
Corruption perception index (10-CPI) 0.005 0.042 0.029 0.045 Test of the differences between Portugal and Italy
V 0.005 0.004 0.035 0.028 Size (log total assets) 0.008 0.005 0.006 0.004
V X DFUxp banks in Spain 0.003 0.002 0.010 0.013 Corruption perception index (10-CPI) 0.235 0.094 0.012 0.014
Market volatility (VIX) 0.131 0.122 0.193 0.281 Test of the differences in the selection equation between Portugal and Italy
V 0.002 0.003 0.003 0.002 Size (log total assets) 0.008 0.032 0.002 0.001
Corruption perception index (10-CPI) 0.005 0.004 0.004 0.003 Test of the differences between Ireland and Spain
V 0.043 0.034 0.013 0.011 Size (log total assets) 0.006 0.005 0.003 0.002
V X DFUxp banks in Spain 0.004 0.002 0.008 0.010 Corruption perception index (10-CPI) 0.001 0.001 0.006 0.004
Market volatility (VIX) 0.028 0.023 0.004 0.006 Test of the differences in the selection equation between Ireland and Spain
V 0.005 0.006 0.002 0.002 Size (log total assets) 0.001 0.001 0.003 0.002
Corruption perception index (10-CPI) 0.002 0.001 0.001 0.001 Test of the differences between Ireland and Italy
V 0.006 0.008 0.013 0.008 Size (log total assets) 0.013 0.016 0.005 0.003
V X DFUxp banks in Spain 0.026 0.031 0.007 0.009 Corruption perception index (10-CPI) 0.018 0.014 0.012 0.013
Market volatility (VIX) 0.026 0.034 0.015 0.019 Test of the differences in the selection equation between Ireland and Italy
V 0.001 0.001 0.001 0.001 Size (log total assets) 0.001 0.001 0.010 0.028
Corruption perception index (10-CPI) 0.003 0.002 0.007 0.006 Test of the differences between Spain and Italy
V 0.026 0.031 0.008 0.029 Size (log total assets) 0.005 0.008 0.004 0.003
V X DFUxp banks in Spain 0.236 0.184 0.026 0.021 Corruption perception index (10-CPI) 0.013 0.010 0.236 0.208
Market volatility (VIX) 0.650 0.635 0.336 0.381 Test of the differences in the selection equation between Spain and Italy V 0.001 0.001 0.001 0.001
Size (log total assets) 0.089 0.136 0.007 0.015 Corruption perception index (10-CPI) 0.012 0.009 0.003 0.002
February 1, 2011
34
TABLE 11 TESTS OF COEFFICIENT DIFFERENCES BETWEEN DMS MODELS ESTIMATED FOR
PRECRISIS YEARS (2003-2006) AND CRISIS YEARS (2007-2008): DFU BANKS RECEIVING STATE AID IN PORTUGAL, IRELAND, SPAIN AND ITALY VS. DFU BANKS IN
THE EUROPEAN SAMPLE The table show the p-values of the tests for coefficient differences as well as the F-test of the overall differences
between the subsamples
Portugal, Ireland, Spain and Italy vs. European sample (PRECRISIS) RV ML RV ML
V 0.014 0.017 0.005 0.006
Size (log total assets) 0.009 0.008 0.128 0.682
V X DFU banks 0.013 0.010 0.016 0.014
Corruption perception index (10-CPI)
0.143 0.138 0.231 0.228
Market volatility (VIX) - - - -
Overall coefficients F-test 0.006 0.010 0.011 0.013
Portugal, Ireland, Spain and Italy vs. European sample (CRISIS) RV ML RV ML
V 0.005 0.004 0.005 0.006
Size (log total assets) 0.008 0.010 0.014 0.026
V X DFU banks 0.403 0.396 0.002 0.012
Corruption perception index (10-CPI)
0.010 0.013 0.036 0.054
Market volatility (VIX) - - - -
Overall coefficients F-test 0.013 0.011 0.014 0.016